Professional Documents
Culture Documents
Systems Engineering
Prof. Olivier L. de Weck
Session 6
Design Definition
Multidisciplinary Optimization
A3 is due today ! A4 is due on Nov 6.
2
3
Multidisclinary Design Optimization (MDO) –
What it is and where it fits in…
CDR
metric 2
Conceptual 4/15
Modeling &
Optimization “Optimal”
metric 1
Design
Decision
Making
metric 3
2/12 Modeling &
Optimization
Concept Concept Concept System
Synthesis variable x1
Screening Selection Design
4
Outline for today
NASA Design Definition Process
Process Overview
5
Design Solution Definition Process
The Design Solution Definition Process is used to translate the outputs of
the Logical Decomposition Process into a design solution definition
6
Design Solution Importance What we wanted
What we got
Cost, performance, schedule
Select Design Solution
Drive down to lowest level
Identify enabling products
7
Design Solution Definition – Best
Practice Process Flow Diagram
Output
Activities
Input
9
Producibility vs. Total Cost
1 bar
2.50 mm
2 bars
0.80 mm
17 bars
More design freedom More complex
(Better performance) (More difficult to optimize)
0.63 mm
10
Concept Question
Which of these three designs
17 bars would you select and why?
8
7
6 2 bars
5 1 bar
4
3
2
1
0
0
0.5
1
1.5
2
2.5
3
Displacement [mm]
11
Outline for today
NASA Design Definition Process
Process Overview
12
Multidisclinary Design Optimization (MDO) –
What it is and where it fits in…
Approximation Concepts
Human Interface
13
MDO - Motivation
© C. W. Miller. All rights reserved. This content is excluded from our Creative Commons
license. For more information, see http://ocw.mit.edu/help/faq-fair-use/.
14
Topic
MDO - Roots
1960 65 70 75 80 85 90 95 2000 2005
MDO Early Years
Schmit's 3 bar truss M MDO roots found in
Gen opt codes appear (Aesop, CONMIN)
LaRC 1st MDO SST papers structural optimization
LaRC IPAD project
LaRC AOO & MDOB & IRO
Government-Sponsored MDO
LaRC SST MDO project
ARC ACSYNT & Applications
EU MOB
NATO AGARD, RTO M M M
Theory, Methods and Frameworks, Tools and Companies
Excel
Matlab
M
M
Optimization algrthms
Mathematica M in mainstream prgms
Integration VRD
Integration Engineous
Integration ALTAIR
Genesis
Integration Phoenix
Concurrent Computing
Linear decomp. M
More complex
Opt Sensit M decomposition
System Sensit M
Approximations techniques appear
Approximation based decomp.
Analytical Target Cascading (Michigan)
Collaborative Optimization (Stanford)
BLISS-LaRC
CSSO-LaRC ND
Visualization UofBuff Commercialization
Commercialization BLISS
Genetic Algorithms
M
of multi-level
Optimality criteria (KKT) algorithm
NASA Glenn NPSS
Physical Programming (RPI)
Isoperformance (MIT)
© Springer. All rights reserved. This content is excluded from our Creative Commons
license. For more information, see http://ocw.mit.edu/help/faq-fair-use/.
Reading: [6a] Agte J., de Weck O., Sobieszczanski-Sobieski J., Arendsen P., Morris A., Spieck M., “MDO: assessment and
direction for advancement - an opinion of one international group”, Structural and Multidisciplinary Optimization, 40 (1), 17
33, January 2010 15
MDO - Example
16
MDO – Method: Bi-Level Integrated System Synthesis
Xsh - Variables
T-throttle
M, h ΛHT- tail sweep
M, h t/c, h, M, ARW, , SREF,SHT,ARHT
LW-wing mom. arm XLOC
LHT-tail mom. arm
h,M,,ARHT,SHT t/c,SHT ,ARW
[t]-thickness array,
ARW,SREF ,,t/c ,SREF,ARHT
size 1x9
[tS]-thickness array,
Propulsion Y^ size 1x9
ESF WE SFC
λ-taper ratio
Xloc={T} Y* D-drag
Y* ESF-eng. scale fact.
Aerodynamics L-lift
Y^ Y^ L Y
D L/D NZ-max. load fact.
R-range
XlocHT,LW,LHT} Y* SFC-spec. fuel cons.
Y* -wing twist
Structures WE-engine weight
W ,
Y^ Y^ WT,WF WF-fuel weight
T
WT-total weight
Xloc={[t],[ts], Y*
ARW- wing aspect ratio
ARHT- tail aspect ratio
WFO,WO h-altitude
H,CDMIN,M<1 Range
NZ M-Mach #
SREF-wing surf. area
SHT-tail surf. area XSH
Y^ t/c-thickness/chord
WBE WFO, WO, NZ, WBE, CDMIN,M<1,H W-wing sweep
R
Constants
17
MDO – Method: Bi-Level Integrated System Synthesis
Structures Range
Xloc-[t],[ts],λ
18
MDO – Method: Bi-Level Integrated System Synthesis
n
f = w1Y^1 + w2Y^2 + w3Y^3 = w i Yi^
i 1 where n = # of Y^ outputs
19
MDO – Method: Bi-Level Integrated System Synthesis
XSH Y* w
20
MDO – Method: Bi-Level Integrated System
Synthesis
System-Level
Optimization
SubSys 1
These make up an approxi-
mated subsystem…
…which is then sent to the
system-level optimization.
21
MDO – Method: Bi-Level Integrated System
Synthesis
System Optimization (SOPT)
Xsh, Y*, w SOPT Formulation
Given: approximation models for opt-
imized subsystem outputs,
minimize: F (Xsh, Y*, w),
SubSys 1 SubSys 2 by varying: Q = {[Xsh],[Y*],[w]}.
Satisfy: c = [Y*]-[Y^] = 0,
Y`s [Xsh,LB] ≤ [Xsh] ≤ [Xsh,UB],
Y^ [Y*LB] ≤ [Y*] ≤ [Y*UB], and
Y*
[wLB] ≤ [w] ≤ [wUB],
Y`s Y`s and retrieve: [Xsh],[Y*],[w], and F at optimum
22
BLISS Cycle # 0
23
BLISS Cycle # 10
24
MDO - Challenges
intermediate
Fidelity Level
fidelity
(e.g. vortex lattice,
beam theory)
can the
empirical results be
models Level of MDO believed?
25
MDO - Challenges
intermediate
Disciplinary Depth
fidelity
(e.g. vortex
lattice, beam
theory)
can the
empirical results be
relations System Breadth believed?
26
Outline for today
NASA Design Definition Process
Process Overview
27
Credit:
28
CDF in industrial setting
ESTEC (ESA)
all of the future projects at ESA are going through the ESA CDF
e.g. CHEOPS
Others
Most NASA centers, ASI, CNES, commercial applications of the
idea (painting, shipbuilding, medical devices)
Benefits
improvements on quality for redesigned products
very quick turnaround for ideas
better cost estimates
increased creativity and productivity in a company
29
Example of Cubesat Design in J-CDS
SwissCube 30
Design of a suborbital space
plane in CDF
31
Requirements
Level 1 requirements.
Reach an altitude of at least 100km over sea level
Zero G-phase flight phase of several minutes
Passenger vehicle carrying 6 people
Level 2 requirements
Safety: load limit 6 g
Spacecraft shall be controllable at any time
Customer experience: view on earth’s curvature and
atmosphere
Environment: The spacecraft’s impact on
32
CDF Design: K1000
33
Requirements verification by modeling
34
Visualization of results.
ALINGHI 2
35
S3 is it feasible? What are the key challenges?
© Swiss Space Systems. All rights reserved. This content is excluded from our Creative
Commons license. For more information, see http://ocw.mit.edu/help/faq-fair-use/.
36
Partner Exercise (5 min)
37
Lessons learned EPFL CDF
38
Lessons learned EPFL CDF (2)
Distributed centers
a lot of information is lost over telecons
videocons are better, but still not ideal, as there is a lot of exchange
near “water cooler”
Staff
pulling people from active projects is problematic
every chair should be at least 2-3-person deep
39
Critical Design Review (CDR)
Critical Design Review (CDR)
Main Purpose: Approve the final design and all its details
Give Green Light to “cut metal” and manufacture the system
Large teams, lots of details …
Can last 1+ week for a large complex project
http://www.techtimes.com/articles/2966/2
0140126/james-webb-space-telescope
passes-last-major-element-level-critical
design-review-eyes-2018-launch.htm
40
CDR Entrance and Success Criteria
42
MIT OpenCourseWare
http://ocw.mit.edu
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.